The traditionally computational methods used by scientists to identify genes involves observing one DNA sequence at a time. Siepel detects two main problems with this approach: “Unfortunately, this method makes it easier to identify faulty replicates of normal genes known as pseudo-genes. Traditional computational methods are also not useful for identifying single exon genes which make up about 10% of human genes. The reason for this is that single exon genes give off fewer signals in their sequence and so are harder to detect by this method, which depends on the signals in the sequence.”

Siepel’s research involves a more successful way of detecting single exon genes: using comparative genomics and evolutionary models. Comparative genomics involves comparing human genomes to that of different organisms. Information about human genomes can be inferred from information gathered from observing simpler organisms.

“Evolutionary models are statistical models that describe the phylogeny or evolution of an organism,” Siepel explains. “It shows the evolutionary branch that led up to that organism.” Observing the phylogeny of humans brings to light the differences that have evolved in their genetic make-up over time. This information helps point scientists to where specific genes may occur.

So far, Siepel’s method of gene detection has been effective; he has an approximate 70 percent success rate. Out of the eight genes he has discovered, the existence of about 200 has been validated by other independent research.